Hi
Anyone else using the Microchip 23LCV512 with battery backup? I'm having trouble with the device apparently locking up between power cycles. In this state it won't respond to read requests and I therefore can't speak about the success or otherwise of writes. I had hoped that moving from prototype to proper PCB would scotch this problem but it remains. I can go for days and many power cycles then suddenly it won't work. I can reliably get it working again by removing both power and battery, powering up then reconnecting the battery. The 'scope shows that attempting a read and triggering on CS results in MISO remaining high-impedance.
In the absence of any solutions can anyone recommend a different device? I only need a few hundred bytes.
Thanks
John
This from Microchip
Hi John,
The pin 3 (NC) is not internally wired, so no issue on it.
I have seen cases before where the command EQIO has been sent to the device, thus the device entered SQI mode (23LCVxxx shares the same die as for the 23LCxxx), and the only workaround was to power down the device (including VBAT pin). The 23LCxxx says to pull up this pin to avoid entering SQI mode, but the 23LCVxxx pin does not have this tied internally to avoid it...
The workaround is the user must ensure that command EQIO command is not sent to the part.
We will release a new family of serial SRAMs that do not have this feature.
Also, if you find that the EQIO command is not sent by mistake, can you please scope Vcc, Vbat and SPI pins during power-up or other conditions where this happens, to understand the environment in which this is happening. The scope plots may point to unusual activity.
Thank you,
....although I wasn't issuing the EQIO command I can confirm that this was what was apparently happening during power-up/down. A pull-up on CS may cure this but it's impossible to prove and I lost faith and went for a FM25040B instead. Plus there's no battery required.
Big points to Microchip for fessing-up.